Design Of Ship Atmospheric Pollutant Monitoring System

With the rapid development of maritime transportation,the problem of atmospheric pollution from ships has become increasingly severe.Currently,research and application of regulatory measures for ship atmospheric pollutant emissions in China are lacking.Due to the unique characteristics of atmospheric pollutants emitted by ships,such as the flow of exhaust plumes,uneven gas concentrations,high temperatures,high levels of water vapor,complex gas components,and harsh monitoring environments,these factors seriously affect the accuracy of ship atmospheric pollutant detection.This thesis is aimed at addressing the difficult problem of ship atmospheric pollutant monitoring by designing Ship Atmospheric Pollutant Monitoring System,which includes low-cost hardware design for the monitoring system,software design for meeting different needs,and algorithm design for improving system monitoring accuracy.The main contents are as follows.（1）By performing distributed sampling of ship stacks,this system solves the problem of low detection accuracy at a single sampling point under non-uniform gas concentrations.The system uses ambient pressure cooling and includes normal pressure type liquid-cooled coils and filters in the gas pre-processing device to address strong interference caused by high water vapor content and over-temperature or damage to gas sensors.The system also features gas calibration,automatic reverse blowing,manual reverse blowing,and other functions to address sensor accuracy deviations resulting from complex gas compositions and prolonged equipment operation,as well pipeline impurities caused by harsh monitoring environments on ships.（2）Using Coati Optimization Algorithm-Gated Recurrent Unit（COA-GRU）to correct the primary data of gas concentration,temperature,humidity,and pressure in each detection room for single point and multiple categories.Subsequently,Kalman Recursive is used to perform multi-point data fusion of the same type on the correction results of gases in each detection chamber.Finally,use the Unscented Kalman Filter algorithm to perform secondary data filtering on the fused data.Two level data processing enables this system to cope with complex detection environments and ensure detection accuracy even under a low-cost sensor architecture.（3）In order to meet different monitoring needs,the following software has been designed:devices display software,upper computer monitoring software,web monitoring software,and mobile application software.devices display software can display monitoring data in real-time.The upper computer monitoring software can not only view static information of ships and atmospheric pollutant emissions,but also remotely control equipment for operations such as reverse blowing and gas calibration.Web monitoring software and mobile application software can facilitate users to remotely view the atmospheric pollutant emissions of ships.The experiment shows that the two-level data processing method proposed in this thesis enables the system to control the indication error of various gas concentrations within±2.5%of the nominal value.Compared with traditional single point sampling monitoring methods,the method proposed in this thesis reduces the indication error by 50%.During ship trials,this system monitored the indication error of various gases with a deviation from the nominal value of the H200D flue gas detection device that complies with national standards of less than 0.5%.In addition,the system’s monitoring accuracy for SO₂,CO,CO₂,and NO₂meets the standard specified in the"General Requirements for Monitoring Emissions of Atmospheric Pollutant from Ships".